Frequently asked questions

How does the program generate trend estimates?

The program uses R to analyze overdispersed Poisson linear models which control for time of day, observer, and point-level variation.

What are the data requirements of the program?

Currently, your point counts must be at least 5 min long and contain at least one distance band. For example, detections recorded as either greater or less than 50m. If the counts are longer that 5 min, detections must have been recorded in time intervals that will allow the data to be reduced to the first 5 min. For example, a 10 minute count where detections are broken into the the first 3, middle 2, and last 5 minutes. Point counts containing only limited distance counts or 10 min counts without any time interval cannot be analyzed with this program.

What if my protocols have changed over time?

As long as all the protocols meet the above requirements (see 'What are the data requirements of the program?'), this is not a problem. Data explorer converts all counts to limited radius 5 min counts so all the data will be comparable.

Why doesn't the program give me estimates of abundance?

While Data Explorer corrects for many potential biases related to detectability (e.g., observer, time of day) it can only calculate relative abundance. There is often not enough information in historical datasets to directly calculate a true abundance. Therefore, just like the breeding bird survey, the program only presents estimates of percent change and not abundance. 

Comments